Antibodies are immunological proteins that bind a specific antigen. In most mammals, including humans and mice, antibodies are constructed from paired heavy and light polypeptide chains. Each chain is made up of individual immunoglobulin (Ig) domains, and thus the generic term immunoglobulin is used for such proteins. Each chain is made up of two distinct regions, referred to as the variable and constant regions. The light and heavy chain variable regions show significant sequence diversity between antibodies, and are responsible for binding the target antigen. The constant regions show less sequence diversity, and are responsible for binding a number of natural proteins to elicit important biochemical events. In humans there are five different classes of antibodies including IgA (which includes subclasses IgA1 and IgA2), IgD, IgE, IgG (which includes subclasses IgG1, IgG2, IgG3, and IgG4), and IgM. The distinguishing features between these antibody classes are their constant regions, although subtler differences may exist in the variable region. FIG. 1 shows an IgG1 antibody, used here as an example to describe the general structural features of immunoglobulins. IgG antibodies are tetrameric proteins composed of two heavy chains and two light chains. The IgG heavy chain is composed of four immunoglobulin domains linked from N- to C-terminus in the order VH-CH1-CH2-CH3, referring to the variable heavy domain, constant heavy domain 1, constant heavy domain 2, and constant heavy domain 3. The IgG CH1, CH2, and CH3 domains are also referred to as constant gamma 1 domain (Cγ1), constant gamma 2 domain (Cγ2), and constant gamma 3 domain (Cγ3) respectively. The IgG light chain is composed of two immunoglobulin domains linked from N- to C-terminus in the order VL-CL, referring to the light chain variable domain and the light chain constant domain respectively.
The variable region of an antibody contains the antigen binding determinants of the molecule, and thus determines the specificity of an antibody for its target antigen. The variable region is so named because it is the most distinct in sequence from other antibodies within the same class. The majority of sequence variability occurs in the complementarity determining regions (CDRs). There are 6 CDRs total, three each per heavy and light chain, designated VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and VL CDR3. The variable region outside of the CDRs is referred to as the framework (FR) region. Although not as diverse as the CDRs, sequence variability does occur in the FR region between different antibodies. Overall, this characteristic architecture of antibodies provides a stable scaffold (the FR region) upon which substantial antigen binding diversity (the CDRs) can be explored by the immune system to obtain specificity for a broad array of antigens. A number of high-resolution structures are available for a variety of variable region fragments from different organisms, some unbound and some in complex with antigen. The sequence and structural features of antibody variable regions are well characterized (Morea et al., 1997, Biophys Chem 68:9-16; Morea et al., 2000, Methods 20:267-279; incorporated by reference), and the conserved features of antibodies have enabled the development of a wealth of antibody engineering techniques (Maynard et al., 2000, Annu Rev Biomed Eng 2:339-376, incorporated by reference). For example, it is possible to graft the CDRs from one antibody, for example a murine antibody, onto the framework region of another antibody, for example a human antibody. This process, referred to in the art as “humanization”, enables generation of less immunogenic antibody therapeutics from nonhuman antibodies. Fragments comprising the variable region can exist in the absence of other regions of the antibody, including for example the antigen binding fragment (Fab) comprising VH-Cγ1 and VH-CL, the variable fragment (Fv) comprising VH and VL, the single chain variable fragment (scFv) comprising VH and VL linked together in the same chain, as well as a variety of other variable region fragments (Little et al., 2000, Immunol Today 21:364-370, incorporated by reference).
The Fc region of an antibody interacts with a number of Fc receptors and ligands, imparting an array of important functional capabilities referred to as effector functions. For IgG the Fc region, as shown in FIG. 1, comprises Ig domains Cγ2 and Cγ3 and the N-terminal hinge leading into Cγ2. An important family of Fc receptors for the IgG class are the Fc gamma receptors (FcγRs). These receptors mediate communication between antibodies and the cellular arm of the immune system (Raghavan et al., 1996, Annu Rev Cell Dev Biol 12:181-220; Ravetch et al., 2001, Annu Rev Immunol 19:275-290). In humans this protein family includes FcγRI (CD64), including isoforms FcγRIa, FcγRIb, and FcγRIc; FcγRII (CD32), including isoforms FcγRIIa (including allotypes H131 and R131), FcγRIIb (including FcγRIIb-1 and FcγRIIb-2), and FcγRIIc; and FcγRIII (CD16), including isoforms FcγRIIIa (including allotypes V158 and F158) and FcγRIIIb (including allotypes FcγRIIIb-NA1 and FcγRIIIb-NA2) (Jefferis et al., 2002, Immunol Lett 82:57-65, incorporated by reference). These receptors typically have an extracellular domain that mediates binding to Fc, a membrane spanning region, and an intracellular domain that may mediate some signaling event within the cell. These receptors are expressed in a variety of immune cells including monocytes, macrophages, neutrophils, dendritic cells, eosinophils, mast cells, platelets, B cells, large granular lymphocytes, Langerhans' cells, natural killer (NK) cells, and γδ T cells. Formation of the Fc/FcγR complex recruits these effector cells to sites of bound antigen, typically resulting in signaling events within the cells and important subsequent immune responses such as release of inflammation mediators, B cell activation, endocytosis, phagocytosis, and cytotoxic attack. The ability to mediate cytotoxic and phagocytic effector functions is a potential mechanism by which antibodies destroy targeted cells. The cell-mediated reaction wherein nonspecific cytotoxic cells that express FcγRs recognize bound antibody on a target cell and subsequently cause lysis of the target cell is referred to as antibody dependent cell-mediated cytotoxicity (ADCC) (Raghavan et al., 1996, Annu Rev Cell Dev Biol 12:181-220; Ghetie et al., 2000, Annu Rev Immunol 18:739-766; Ravetch et al., 2001, Annu Rev Immunol 19:275-290, incorporated by reference). The cell-mediated reaction wherein nonspecific cytotoxic cells that express FcγRs recognize bound antibody on a target cell and subsequently cause phagocytosis of the target cell is referred to as antibody dependent cell-mediated phagocytosis (ADCP). A number of structures have been solved of the extracellular domains of human FcγRs, including FcγRIIa (pdb accession code 1H9V) (Sondermann et al., 2001, J Mol Biol 309:737-749) (pdb accession code 1FCG) (Maxwell et al., 1999, Nat Struct Biol 6:437-442), FcγRIIb (pdb accession code 2FCB) (Sondermann et al., 1999, Embo J 18:1095-1103); and FcγRIIIb (pdb accession code 1E4J) (Sondermann et al., 2000, Nature 406:267-273, incorporated by reference). All FcγRs bind the same region on Fc, at the N-terminal end of the Cγ2 domain and the preceding hinge, shown in FIG. 2. This interaction is well characterized structurally (Sondermann et al., 2001, J Mol Biol 309:737-749 incorporated by reference), and several structures of the human Fc bound to the extracellular domain of human FcγRIIIb have been solved (pdb accession code 1E4K) (Sondermann et al., 2000, Nature 406:267-273) (pdb accession codes 1IIS and 1IIX) (Radaev et al., 2001, J Biol Chem 276:16469-16477, incorporated by reference), as well as has the structure of the human IgE Fc/FcεRIa complex (pdb accession code 1F6A) (Garman et al., 2000, Nature 406:259-266, incorporated by reference).
The different IgG subclasses have different affinities for the FcγRs, with IgG1 and IgG3 typically binding substantially better to the receptors than IgG2 and IgG4. All FcγRs bind the same region on IgG Fc, yet with different affinities: the high affinity binder FcγRI has a Kd for IgG1 of 10−8M−1, whereas the low affinity receptors FcγRII and FcγRIII generally bind at 10−6 and 10−5 respectively. The extracellular domains of FcγRIIIa and FcγRIIIb are 96% identical, however FcγRIIIb does not have a intracellular signaling domain. Furthermore, whereas FcγRI, FcγRIIa/c, and FcγRIIIa are positive regulators of immune complex-triggered activation, characterized by having an intracellular domain that has an immunoreceptor tyrosine-based activation motif (ITAM), FcγRIIb has an immunoreceptor tyrosine-based inhibition motif (ITIM) and is therefore inhibitory. Thus the former are referred to as activation receptors, and FcγRIIb is referred to as an inhibitory receptor. The receptors also differ in expression pattern and levels on different immune cells. Yet another level of complexity is the existence of a number of FcγR polymorphisms in the human proteome. A particularly relevant polymorphism with clinical significance is V158/F158 FcγRIIIa. Human IgG1 binds with greater affinity to the V158 allotype than to the F158 allotype. This difference in affinity, and presumably its effect on ADCC and/or ADCP, has been shown to be a significant determinant of the efficacy of the anti-CD20 antibody rituximab (Rituxan®), a registered trademark of IDEC Pharmaceuticals Corporation). Patients with the V158 allotype respond favorably to rituximab treatment; however, patients with the lower affinity F158 allotype respond poorly (Cartron et al., 2002, Blood 99:754-758, incorporated by reference). Approximately 10-20% of humans are V158/V158 homozygous, 45% are V158/F158 heterozygous, and 35-45% of humans are F158/F158 homozygous (Lehrnbecher et al., 1999, Blood 94:4220-4232; Cartron et al., 2002, Blood 99:754-758, incorporated by reference). Thus 80-90% of humans are poor responders, that is they have at least one allele of the F158 FcγRIIIa.
An overlapping but separate site on Fc, shown in FIG. 1, serves as the interface for the complement protein C1q. In the same way that Fc/FcγR binding mediates ADCC, Fc/C1q binding mediates complement dependent cytotoxicity (CDC). C1q forms a complex with the serine proteases C1r and C1s to form the C1 complex. C1q is capable of binding six antibodies, although binding to two IgGs is sufficient to activate the complement cascade. Similar to Fc interaction with FcγRs, different IgG subclasses have different affinity for C1q, with IgG1 and IgG3 typically binding substantially better to the FcγRs than IgG2 and IgG4. There is currently no structure available for the Fc/C1q complex; however, mutagenesis studies have mapped the binding site on human IgG for C1q to a region involving residues D270, K322, K326, P329, and P331, and E333 (Idusogie et al., 2000, J Immunol 164:4178-4184; Idusogie et al., 2001, J Immunol 166:2571-2575, incorporated by reference).
A site on Fc between the Cγ2 and Cγ3 domains, shown in FIG. 1, mediates interaction with the neonatal receptor FcRn, the binding of which recycles endocytosed antibody from the endosome back to the bloodstream (Raghavan et al., 1996, Annu Rev Cell Dev Biol 12:181-220; Ghetie et al., 2000, Annu Rev Immunol 18:739-766, incorporated by reference). This process, coupled with preclusion of kidney filtration due to the large size of the full length molecule, results in favorable antibody serum half-lives ranging from one to three weeks. Binding of Fc to FcRn also plays a key role in antibody transport. The binding site for FcRn on Fc is also the site at which the bacterial proteins A and G bind. The tight binding by these proteins is typically exploited as a means to purify antibodies by employing protein A or protein G affinity chromatography during protein purification. Thus the fidelity of this region on Fc is important for both the clinical properties of antibodies and their purification. Available structures of the rat Fc/FcRn complex (Martin et al., 2001, Mol Cell 7:867-877, incorporated by reference), and of the complexes of Fc with proteins A and G (Deisenhofer, 1981, Biochemistry 20:2361-2370; Sauer-Eriksson et al., 1995, Structure 3:265-278; Tashiro et al, 1995, Curr Opin Struct Biol 5:471-481, incorporated by reference) provide insight into the interaction of Fc with these proteins.
A key feature of the Fc region is the conserved N-linked glycosylation that occurs at N297, shown in FIG. 1. This carbohydrate, or oligosaccharide as it is sometimes referred, plays a critical structural and functional role for the antibody, and is one of the principle reasons that antibodies must be produced using mammalian expression systems. While not wanting to be limited to one theory, it is believed that the structural purpose of this carbohydrate may be to stabilize or solubilize Fc, determine a specific angle or level of flexibility between the Cγ3 and Cγ2 domains, keep the two Cγ2 domains from aggregating with one another across the central axis, or a combination of these. Efficient Fc binding to FcγR and C1q requires this modification, and alterations in the composition of the N297 carbohydrate or its elimination affect binding to these proteins (Umaña et al., 1999, Nat Biotechnol 17:176-180; Davies et al., 2001, Biotechnol Bioeng 74:288-294; Mimura et al., 2001, J Biol Chem 276:45539-45547; Radaev et al., 2001, J Biol Chem 276:16478-16483; Shields et al., 2001, J Biol Chem 276:6591-6604; Shields et al., 2002, J Biol Chem 277:26733-26740; Simmons et al., 2002, J Immunol Methods 263:133-147, incorporated by reference). Yet the carbohydrate makes little if any specific contact with FcγRs (Radaev et al., 2001, J Biol Chem 276:16469-16477, incorporated by reference), indicating that the functional role of the N297 carbohydrate in mediating Fc/FcγR binding may be via the structural role it plays in determining the Fc conformation. This is supported by a collection of crystal structures of four different Fc glycoforms, which show that the composition of the oligosaccharide impacts the conformation of Cγ2 and as a result the Fc/FcγR interface (Krapp et al., 2003, J Mol Biol 325:979-989, incorporated by reference).
The features of antibodies discussed above—specificity for target, ability to mediate immune effector mechanisms, and long half-life in serum—make antibodies powerful therapeutics. Monoclonal antibodies are used therapeutically for the treatment of a variety of conditions including cancer, inflammation, and cardiovascular disease. There are currently over ten antibody products on the market and hundreds in development. In addition to antibodies, an antibody-like protein that is finding an expanding role in research and therapy is the Fc fusion (Chamow et al., 1996, Trends Biotechnol 14:52-60; Ashkenazi et al., 1997, Curr Opin Immunol 9:195-200, incorporated by reference). An Fc fusion is a protein wherein one or more polypeptides is operably linked to Fc. An Fc fusion combines the Fc region of an antibody, and thus its favorable effector functions and pharmacokinetics, with the target-binding region of a receptor, ligand, or some other protein or protein domain. The role of the latter is to mediate target recognition, and thus it is functionally analogous to the antibody variable region. Because of the structural and functional overlap of Fc fusions with antibodies, the discussion on antibodies in the present invention extends directly to Fc fusions.
Uchida et al. (J Exp Med v 199, p 1659, 21 June 2004) found that anti-CD20 monoclonal antibody depletion of B cells in mice is greatly impaired in FcR common gamma chain knockouts. This impairment is 60-70% for blood B cells and is complete for splenic B cells. Depletion is not significantly impaired in their model for FcγRII or FcγRIII knockouts, and is only somewhat depleted in FcγRI knockouts, showing that either activating receptor RIII or RI is sufficient. It is noted that macrophages possess both RI and RIII, whereas natural killer cells possess only RIII. These findings and further experiments with this model in genetic and pharmacological knockouts of specific effector cell lineages demonstrates that macrophage activity was the impaired component that caused loss of anti-CD20 cytolysis, and that neither complement nor NK cells were responsible for this abrogation. Taken together these results provide strong support for the hypothesis that macrophages are a key effector cell type for anti-CD20 therapy and that depletion of B cells likely involves BOTH phagocytic and secreted cytolytic factors. Also, depletion of splenic B-cells, clearly less rapid and requiring more potency than blood B cells, is achieved by anti-CD20 antibodies that effectively recruit macrophages, and splenic B cell depletion is greatly impaired by disruption of macrophage activity or FcR common gamma chain knockout.
There are a number of possible mechanisms by which antibodies destroy tumor cells, including anti-proliferation via blockage of needed growth pathways, intracellular signaling leading to apoptosis, enhanced down regulation and/or turnover of receptors, CDC, ADCC, ADCP, and promotion of an adaptive immune response (Cragg et al., 1999, Curr Opin Immunol 11:541-547; Glennie et al., 2000, Immunol Today 21:403-410. incorporated by reference). Anti-tumor efficacy may be due to a combination of these mechanisms, and their relative importance in clinical therapy appears to be cancer dependent. Despite this arsenal of anti-tumor weapons, the potency of antibodies as anti-cancer agents is unsatisfactory, particularly given their high cost. Patient tumor response data show that monoclonal antibodies provide only a small improvement in therapeutic success over normal single-agent cytotoxic chemotherapeutics. For example, just half of all relapsed low-grade non-Hodgkin's lymphoma patients respond to the anti-CD20 antibody rituximab (McLaughlin et al., 1998, J Clin Oncol 16:2825-2833. incorporated by reference). Of 166 clinical patients, 6% showed a complete response and 42% showed a partial response, with median response duration of approximately 12 months. Trastuzumab (Herceptin®, a registered trademark of Genentech), an anti-HER2/neu antibody for treatment of metastatic breast cancer, has less efficacy. The overall response rate using trastuzumab for the 222 patients tested was only 15%, with 8 complete and 26 partial responses and a median response duration and survival of 9 to 13 months (Cobleigh et al., 1999, J Clin Oncol 17:2639-2648, incorporated by reference). Currently for anticancer therapy, any small improvement in mortality rate defines success. Thus there is a significant need to enhance the capacity of antibodies to destroy targeted cancer cells.
The role of FcγR-mediated effector functions in the anti-cancer activity of antibodies has been demonstrated in mice (Clynes et al., 1998, Proc Natl Acad Sci USA 95:652-656; Clynes et al., 2000, Nat Med 6:443-446, incorporated by reference), and the affinity of interaction between Fc and certain FcγRs correlates with targeted cytotoxicity in cell-based assays (Shields et al., 2001, J Biol Chem 276:6591-6604; Presta et al., 2002, Biochem Soc Trans 30:487-490; Shields et al., 2002, J Biol Chem 277:26733-26740, incorporated by reference). Additionally, a correlation has been observed between clinical efficacy in humans and their allotype of high (V158) or low (F158) affinity polymorphic forms of FcγRIIIa (Cartron et al., 2002, Blood 99:754-758, incorporated by reference).
Mutagenesis studies have been carried out on Fc towards various goals, with substitutions typically made to alanine (referred to as alanine scanning) or guided by sequence homology substitutions (Duncan et al., 1988, Nature 332:563-564; Lund et al., 1991, J Immunol 147:2657-2662; Lund et al., 1992, Mol Immunol 29:53-59; Jefferis et al., 1995, Immunol Lett 44:111-117; Lund et al., 1995, Faseb J 9:115-119; Jefferis et al., 1996, Immunol Lett 54:101-104; Lund et al., 1996, J Immunol 157:4963-4969; Armour et al., 1999, Eur J Immunol 29:2613-2624; Shields et al., 2001, J Biol Chem 276:6591-6604) (U.S. Pat. No. 5,624,821; U.S. Pat. No. 5,885,573; PCT WO 00/42072; PCT WO 99/58572), all incorporated by reference. The majority of substitutions reduce or ablate binding with FcγRs. However some success has been achieved at obtaining Fc variants with higher FcγR affinity. (See for example U.S. Pat. No. 5,624,821 and PCT WO 00/42072). For example, Winter and colleagues substituted the human amino acid at position 235 of mouse IgG2b antibody (a glutamic acid to leucine mutation) that increased binding of the mouse antibody to human FcγRI by 100-fold (Duncan et al., 1988, Nature 332:563-564) (U.S. Pat. No. 5,624,821). Shields et al. used alanine scanning mutagenesis to map Fc residues important to FcγR binding, followed by substitution of select residues with non-alanine mutations (Shields et al., 2001, J Biol Chem 276:6591-6604; Presta et al., 2002, Biochem Soc Trans 30:487-490) (PCT WO 00/42072), incorporated by reference.
Enhanced affinity of Fc for FcγR has also been achieved using engineered glycoforms generated by expression of antibodies in engineered or variant cell lines (Umaña et al., 1999, Nat Biotechnol 17:176-180; Davies et al., 2001, Biotechnol Bioeng 74:288-294; Shields et al., 2002, J Biol Chem 277:26733-26740; Shinkawa et al., 2003, J Biol Chem 278:3466-3473, incorporated by reference). This approach has generated enhancement of the capacity of antibodies to bind FcγRIIIa and to mediate ADCC.
A substantial obstacle to engineering anti-CD20 antibodies with the desired properties is the difficulty in predicting what amino acid modifications, out of the enormous number of possibilities, will achieve the desired goals, coupled with the inefficient production and screening methods for antibodies. Indeed one of the principle reasons for the incomplete success of the prior art is that approaches to Fc engineering have thus far involved hit-or-miss methods such as alanine scans or production of glycoforms using different expression strains.
In summary, there is a need for antibodies with enhanced therapeutic properties. Despite such widespread use, anti-CD20 antibodies are not optimized for clinical use. Two significant deficiencies of antibodies are their suboptimal anticancer potency and their demanding production requirements. In these studies, the Fc modifications that were made were fully or partly random in hopes of obtaining variants with favorable properties. These deficiencies are addressed by the present invention.